High-Resolution Spatial Mapping of Electrocorticographic Activities with a 4096-Channel, Multiplexed Thin-Film Transistor Array

Advanced brain-machine interfaces (BMIs) demand implantable devices that offer flexibility, high-density and high-throughput capabilities. Conventional neural electrodes are constrained by the wiring strategy. Here we present a flexible and implantable electrocorticography (ECoG) device array (NeuroCam) based on metal-oxide semiconductor thin-film transistors (TFTs), to record brain activities at a large scale. Employing a multiplexing technique, the system is capable to record ECoG signals with up to 4096 channels and a density of 44 sites/mm 2 , while compressing the fan-in/fan-out leads to around a hundred. In a rabbit model with epilepsy, the NeuroCam array maps abnormal spike-wave discharges with an exceptional spatial resolution (150 μm) across extensive brain areas. The device strategy provides a promising route towards high-throughput BMIs with potential applications in fundamental neuroscience studies and practical biomedicine. Significance Statement Developing high-throughput neural recording techniques is critical for next-generation brain-machine interfaces (BMIs), but conventional neural electrodes are limited by the difficulty of scaling the lead wires. Here we report an implantable neural interface, based on a flexible thin-film transistor (TFT) array named NeuroCam. Employing an actively switching (multiplexing) technique, NeuroCam is capable to record electrocorticographic (ECoG) signals with up to 4096 channels and a density of 44 sites/mm 2 , with only 128 input/output leads. Validated in a rabbit model of epilepsy, NeuroCam records spatially resolved neural activities across extensive cortical areas with a resolution of 150 μm, and precisely maps the temporal evolution of epileptic waves. NeuroCam shows a promising solution for high-performance BMIs.